search

TixiaoShan / LIO-SAM

LIO-SAM: Tightly-coupled Lidar Inertial Odometry via Smoothing and Mapping
BSD 3-Clause "New" or "Revised" License
3.34k stars 1.24k forks source link

Run big-loop.bag data set with different results #21

Closed windwolf9 closed 4 years ago

windwolf9 commented 4 years ago

Hi,@TixiaoShan,I run big-loop.bag dataset, the results of loop closure is different，and Sometimes the results was good, sometimes it was bad! (Three times out of ten the results were good) Did I miss some key steps？ Thanks!

The results are as follows：

20200714181842

20200714182539

TixiaoShan commented 4 years ago

The loop closure results come from PCL icp. Sometimes ICP gives bad results. A more advanced loop closure can be found here. https://github.com/irapkaist/SC-LeGO-LOAM

windwolf9 commented 4 years ago

Thanks,and I found that the loop closure is good In the beginning()，and it will going to get worse immediately and an error occurs :[ WARN] [1594800092.012221229]: Large velocity, reset IMU-preintegration! When it happened，then the maps got worse。 The results are as follows： 20200715162417 20200715162513 20200715162557

TixiaoShan commented 4 years ago

The current implementation of loop closure is not ideal. The success rate heavily depends on the performance of PCL icp.

TixiaoShan commented 4 years ago

A better implementation is welcome from the community. Feel free to join and discuss how to improve its performance.

TixiaoShan commented 4 years ago

A better implementation is welcome from the community. Feel free to join and discuss how to improve its performance.

Loop closure should be fixed now.